In this paper, we proposed an all-sapphire-based extrinsic Fabry-Perot interferometer (EFPI) pressure sensor based on an optimized wet etching process, aiming to improve the quality of the interference signal. The sapphire pressure sensitive diaphragm (SPSD) was fabricated by wet etching solutions with different mixture ratios of H3PO4 and H2SO4 at 280°C. The differences of mixture ratios affect the surface roughness of SPSD. SPSDs with surface roughness of 3.91nm and 0.39nm are obtained when the mixture ratios of H3PO4 and H2SO4 is 1:1 and 1:3, respectively. We constructed pressure sensing test system adopting these two kinds of SPSD and performed comparative test. The experiment results show that the demodulation jump can be solved and cavity length fluctuation is decreased to ±5nm when the surface roughness of SPSD is 0.39nm.
© 2021 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
Accurate measurement of pressure in high-temperatures environment (above 1000 °C) is of great significance in turbine engines, high-speed combustors, and other aerospace applications . Conventional electronic pressure sensors fabricated using MEMS technology have good performance at general temperature. However, due to the plastic deformation of silicon materials of about 520°C - 600°C  and the temperature limitation of silicon electronics , this kind of sensor cannot work reliably in harsh environment of high-temperature for long time. It is generally considered that conventional silicon-based electronics typically cannot survive extended operation beyond 250°C . For example, silicon-based piezoresistive pressure sensor with integrated signal-conditioning circuit can only operate for a long time in the temperature below 220°C . Compared with silicon-based sensors, the SiC pressure sensors based on piezoresistive and capacitive detection principles have obvious advantages at high working temperatures up to 600°C . However, the sensing performance of electromechanical-based pressure sensors at ultra-high temperature  is limited by the silicon diaphragm and the underlying electrode. The application of electromechanical-based pressure sensors in high-temperature environments above 1000℃ is still restricted.
In recent years, fiber-optic sensors have great potential in high-temperature applications . Compared to silicon and SiC-based high-temperature pressure sensors, fiber-optic pressure sensors are modulated in the optical signal by the sensitive unit. The signal is transmitted to the remote end through the optical fiber for demodulation. Thus, fiber-optical pressure sensor is no longer affected by the deterioration of the electrical performance in the high temperature environment. On the other hand, it has the advantages of high temperature resistance, electromagnetic interference immunity, corrosion and oxidation resistance [8,9]. Diaphragm-based EFPI pressure sensors based on different materials of sensitive unit have been widely used in pressure measurement fields [10,11] and the working temperature of the sensor depends on the materials of sensitive unit. For example, the silica-based EFPI pressure sensors are ultimately restricted by the softening of the glass diaphragm, which begin to occur below 1000°C . Therefore, for advanced optical sensors, it is necessary to use high temperature resistant materials in order to avoid the thermal limit of silica glass.
Sapphire is an excellent crystal with high temperature resistance and wide transmission spectral range. It has high melting point up to 2050°C, which makes it an ideal material for high temperature sensing . Jihaeng Yi et al.  proposed an all-sapphire EFPI pressure sensor with a sapphire substrate and a sapphire pressure sensitive diaphragm made by dry etching process, which is bonded to the sensitive unit. Dry etching has high anisotropy and faster etching rate, but the substrate is easily damaged and the cost is relatively high . Compared with dry etching, wet etching has advantages of faster etching rate along a certain crystal orientation, less damage to the material surface, low cost and high production efficiency. However, technological challenge remains for sapphire EFPI pressure sensors based on wet etching process, such as roughness control of SPSD, which determines the quality of the interference signal of the sensor and would further affect the accuracy and stability of the demodulation results.
In this paper, we proposed an all-sapphire-based EFPI pressure sensor based on wet etching, and optimized the mixture ratios of wet etching solutions to reach a higher signal quality. The control of surface roughness of SPSD by different mixture ratios of H3PO4 and H2SO4 was analyzed and a 0.39nm surface roughness was obtained. We carried out pressure experiment to verify the performance of the proposed sensor. The experimental results indicate that the proposed pressure sensor shows a good linearity in the pressure range of 0∼5MPa, with a resolution of 0.3% F.S., which solves the problem of demodulation jump and reduces the demodulation fluctuation.
2. Sensor fabrication and principle of operation
The fabrication procedure of wet anisotropic etching discussed in this paper is shown in Fig. 1. A C-plane (0001) highly polished sapphire wafer obtained commercially is used. The thickness of the sapphire wafers is 210µm and the root mean square roughness measured by 3D confocal microscope is ∼0.33 nm. The SiO2 films of 20µm thickness are deposited on the sapphire wafer by plasma-enhanced chemical vapor deposition (PECVD). A 2mm×2 mm square is patterned on the SiO2 films by photolithography, as shown in Fig. 1(c). Subsequently, SiO2 is etched by reactive ion etching (RIE) to obtain a SiO2 mask, and the photoresist is removed by means of wet etching using remover. After the etching of the sapphire, SiO2 mask is removed by HF solutions and graphical sapphire diaphragm is obtained.
The surface roughness of the etched sapphire diaphragm directly affects the performance of the sensor. Because the roughness of the etched surface obtained by dry etching is poorer than that of wet etching, here we use wet etching method with a mixture solution of H3PO4 and H2SO4. The surface roughness of the sapphire diaphragm corroded by the different proportions of the etching solution is analyzed in order to determine the optimal proportion concentration. Mixture solution of H3PO4 and H2SO4 with the volume ratio of 1:1, 1:2 and 1:3 is used respectively for wet etching at 280°C and 100µm etching cavity as the fabrication target. Through the finite element analysis, 100µm deep etching is suitable for the diaphragm to maintain good mechanical properties in the pressure range of 0 MPa to 5 MPa. We found that with the proportion of H2SO4 increasing in the mixed solution, the etching rate is 2.04µm/min, 1.53µm/min and 1.04µm/min and the etching time is 50minutes, 65minutes and 96minutes, respectively. The etching roughness of sapphire diaphragm after wet etching is tested by using 3D confocal microscope. The surface metrology of SPSD samples after wet etching are shown in Fig. 2. When the volume ratio of H3PO4 and H2SO4 are 1:1, 1:2 and 1:3, the etching roughness of SPSD is 3.91 nm, 2.17 nm and 0.39 nm, respectively. From the surface metrology, we discover that the etching roughness of SPSD decreases as the volume ratio of H3PO4 and H2SO4 decreases. Because the decrease of roughness of SPSD is beneficial to the improvement of the signal quality of sensor, we bond the sapphire diaphragm etched by the mixture solution of H3PO4 and H2SO4 with the volume ratio of 1:3 with sapphire substrate to achieve sapphire EFPI pressure sensing.
The sapphire pressure sensitive chip contains sapphire diaphragm and sapphire substrate. The bonding process of the chip consists of sapphire pre-bonding and sapphire direct bonding. Before pre-bonding, surface treatment is carried out to remove any oxide remaining on surfaces of sapphire substrates. Then, sapphire diaphragm and sapphire substrate are pattern-aligned and mounted on clamping vise to carry out pre-bonding under 200°C in vacuum environment for 1 h. To form sapphire sensitive structure and strengthen the bonding, the pre-bonding sapphire chip is transferred to high-temperature vacuum bonding furnace for direct bonding, in which the sapphire chip is baked at 1250°C for 45 h. Figure 3 shows the schematic diagram of the sapphire pressure sensor. The upper surface of the sapphire substrate and the lower surface of the sapphire diaphragm constitute the pressure-sensitive F-P cavity. In Fig. 3, we use angled physical contact type single mode optical fiber as the carrier of signal transmission to eliminate the potential F-P cavity structure between the optical fiber and the sapphire substrate. In order to facilitate the installation of the sensor, we roughen the upper surface of the sapphire diaphragm, so only the upper and lower surfaces of the sapphire substrate and the lower surface of the sapphire diaphragm produce reflected light. When external pressure acts on the sensitive diaphragm, the sensitive diaphragm is deformed, and the deformation ω of the sensitive diaphragm is expressed as:
In order to investigate the effect of etching roughness of SPSD on the performance of EFPI pressure sensing system, the two kinds of SPSDs are bonded with a 430µm thickness sapphire substrate respectively to form sapphire pressure sensitive chip. The EFPI pressure sensing test system using the two kinds of sapphire pressure sensitive chips is constructed and contrast test is performed. Figure 4 shows the schematic diagram of EFPI pressure sensing test system. The broadband light from optical source is launched into a circulator and then injected into the center of the sapphire pressure sensitive chip. After the two reflected lights interfere with each other, they are transmitted to the spectrum analysis module through fiber circulator, and the spectral data is transmitted to the computer through the USB interface for data analysis to realize the demodulation of the pressure. In order to ensure the sealing performance, the sapphire pressure sensitive chip is packaged with high-temperature alloy material by elastic sealing technique.
The light emitted from the light source is incident on the sapphire pressure sensitive chips through the circulator and optical fiber. The Fresnel reflection of sapphire-air reflecting surface of the sapphire pressure sensitive chip is approximately 7.6%. Due to the low reflectance of the sapphire-air surface, advanced reflection can be ignored, so the total reflected light intensity of the three reflecting surfaces in the F-P sensor (shown in Fig. 3) can be expressed as :
A simplified form after filtering out DC terms and select the interference signal of the pressure-sensitive F-P cavity can be expressed as:5. From the interference spectrum, we discover that the contrast of interference spectrum increases with the decreasing of surface roughness of sapphire pressure sensitive diaphragm. The reflectivity of the optical plane with 0.39 nm etching roughness is significantly higher than that of the optical plane with 3.19 nm etching roughness. After data analysis, the contrast of fringe with 0.39 nm and 3.91 nm etching roughness are 0.385 and 0.114 respectively. This means that higher optical signal-to-noise ratio can tolerate more signal transmission errors, reduce the accuracy requirements of optical sensor system devices and system construction cost and the construction cost of the whole system. The service time of the sensor system is prolonged and the stable pressure measurement is guaranteed.
To extract pressure-sensitive F-P cavity interference signal from interference spectrum, the discrete cross-correlation operation is used to demodulate the spectral information. The discrete cross-correlation operation numerically simulates low coherence interference spectrums with different optical path differences (OPDs), and performs discrete cross-correlation operation between the simulated interference signals and the interference signals obtained from actual sampling, and calculates the correlation coefficient. By continuously changing the value of the simulated OPD, the simulated low coherence interference spectrum changes, thus the calculated correlation number can be seen as a function of the simulated OPD. The maximum correlation value is obtained when the simulated OPD is equal to the OPD of the measured fiber interferometer. Therefore, the cavity length value of the virtual normalized interference spectrum signal with the largest correlation coefficient is identified as the cavity length value of the actual F-P sensor.
A simplified and normalized interference spectrum of a virtual F-P fiber sensor with cavity length L can be constructed as:
During the experiment, the pressurization precision of the pressure sensor test system is precisely controlled by pressure controller, and the pressure controller provides pressure input through a high-purity nitrogen cylinder. The sensor is connected with the pressure controller by pressure pipeline. The output pressure value of the pressure controller is maintained at 1 MPa, and the interference spectrum is collected every 30s. A total of 100 interference spectrums are collected. We perform discrete cross-correlation operation to every interference spectrum respectively. Figure 6 shows the maximum value positions of discrete cross-correlation function of different etching roughness of sapphire pressure sensitive chips. It can be seen that jump error does not occur when the etching roughness is 0.39 nm, while 10 jump errors occurs when the etching roughness is 3.91 nm.
It can also be seen that as the surface roughness decreases, the demodulation fluctuation becomes smaller. The maximum value position of discrete cross-correlation function is stable at ±5 nm when the surface roughness is 0.39 nm. The fluctuation of the cavity length is approximately 0.3%F.S. (full scale). In order to investigate the dynamic sensor response, we perform pressure experiment to the SPSD of 0.39 nm etching roughness. The output pressure of pressure controller increases and decreases over a range of 0∼5 MPa at 1 MPa intervals, as shown in Fig. 7. Within the pressure test range, the output optical path of sensor system decreases linearly with the increase of the applied pressure. A calibration curve is fitted to the sensor data via a quadratic regression (R2=1) and is compared with data from an electronic pressure gauge.
In conclusion, this paper study the effects of mixture ratios of wet etching solution on the surface roughness of SPSD. We find that the surface roughness of SPSD decreases as the volume ratio of H3PO4 and H2SO4 decrease. Through optimize the mixture ratios of wet etching solution, the SPSD with 0.39 nm surface roughness is obtained. The experiment results show that demodulation jump is solved and cavity length fluctuation is decreased to ±5 nm when the surface roughness of SPSD is 0.39 nm. The pressure sensor shows a good linearity in the range of 0≃5 MPa, with a resolution of 0.3% F.S. As the all-sapphire pressure sensor needs to work at temperatures over 1200°C, future studies must work to develop sapphire lead-in fiber and high-temperature packaging.
National High-tech Research and Development Program (2015AA043505).
The authors declare no conflicts of interest.
1. W. Pulliam, P. Russler, and R. Fielder, “In High-temperature, high-bandwidth, fiber optic, MEMS pressure-sensor technology for turbine engine component testing,” Proc. SPIE 4578, 229–238 (2002). [CrossRef]
2. R. Juan, M. Ward, P. Kinnell, R. Craddock, and X. Y. Wei, “Plastic Deformation of Micromachined Silicon Diaphragms with a Sealed Cavity at High Temperatures,” Sensors 16(2), 204 (2016). [CrossRef]
3. J. Yang, “A harsh environment wireless pressure sensing solution utilizing high temperature electronics,” Sensors 13(3), 2719–2734 (2013). [CrossRef]
4. Z. Yao, T. Liang, P. G. Jia, Y. P. Hong, L. Qi, C. Lei, B. Zhang, and J. J. Xiong, “A High-Temperature Piezoresistive Pressure Sensor with an Integrated Signal-Conditioning Circuit,” Sensors 16(6), 913 (2016). [CrossRef]
5. R. S. Okojie, “Stable 600 _C silicon carbide MEMS pressure transducers,” Proc. SPIE 6555, 65550V (2007). [CrossRef]
6. N. Marsia, B. Y. Majlisa, A. A. Hamzaha, and F. Mohd-Yasin, “A MEMS packaged capacitive pressure sensor employing 3C-SiC with operating temperature of 500 °C,” Microsyst. Technol. 21(1), 9–20 (2015). [CrossRef]
7. B. T. Wang, Y. X. Niu, S. W. Zheng, Y. H. Yin, and M. Ding, “A high temperature sensor based on sapphire fiber Fabry–Pérot interferometer,” IEEE Photonics Technol. Lett. 32(2), 89–92 (2020). [CrossRef]
8. S. Pevec and D. Donlagic, “All-fiber, long-active-length Fabry-Perot strain sensor,” Opt. Express 19(16), 15641–15651 (2011). [CrossRef]
9. X. L. Zhou, Q. X. Yu, and W. Peng, “Fiber-optic Fabry-Perot pressure sensor for down-hole application,” Opt. Lasers Eng. 121, 289–299 (2019). [CrossRef]
10. W. Parkes, V. Djakov, J. S. Barton, S. Watson, W. N. MacPherson, J. T. M. Stevenson, and C. C. Dunare, “Design and fabrication of dielectric diaphragm pressure sensors for applications to shock wave measurement in air,” J. Micromech. Microeng. 17(7), 1334–1342 (2007). [CrossRef]
11. X. G. Qi, S. Wang, J. F. Jiang, K. Liu, X. Wang, Y. G. Yang, and T. G. Liu, “Fiber Optic Fabry-Perot Pressure Sensor with Embedded MEMS Micro-Cavity for Ultra-High Pressure Detection,” J. Lightwave Technol. 37(11), 2719–2725 (2019). [CrossRef]
12. J. Yi, “Adhesive-Free Bonding of Monolithic Sapphire for Pressure Sensing in Extreme Environments,” Sensors 18(8), 2712 (2018). [CrossRef]
13. S. Yang, D. Homa, G. Pickrell, and A. B. Wang, “Fiber Bragg grating fabricated in micro-single-crystal sapphire fiber,” Opt. Lett. 43(1), 62–65 (2018). [CrossRef]
14. J. Yi, E. Lally, A. B. Wang, and Y. Xu, “Demonstration of an All-Sapphire Fabry–Pérot Cavity for Pressure Sensing,” IEEE Photonic. Tech. L. 23(1), 9–11 (2010). [CrossRef]
15. Y. Q. Shang, H. Qi, Y. L. Ma, Y. L. Wu, Y. Zhang, and J. Chen, “Study on sapphire microstructure processing technology based on wet etching,” Int. J. Mod. Phys. B 31(07), 1741004 (2017). [CrossRef]
16. Z. Wang, J. Chen, H. Wei, H. Liu, Z. Ma, N. Chen, Z. Chen, T. Wang, and F. Pang, “Sapphire Fabry–Perot interferometer for high-temperature pressure sensing,” Appl. Opt. 59(17), 5189–5196 (2020). [CrossRef]